FOREIGN AID ON POVERTY REDUCTION IN EMPIRICAL STUDY: KENYA

Jacob Oduor and Dickson KhaingaGDN Working Paper No. 34

Sharkhuu MunkhbatAid and Development03.12.2011

2

CONTENTS

Introduction Trends in foreign aid and poverty Empirical strategy Empirical model - Difference-in-

difference Empirical Results - When impact is felt only in the month of

disbursement - When impact is felt in the month of

disbursement and thereafter - Impact of ODA after the 2005 Paris Declaration Conclusions

3

INTRODUCTION

In Kenya, the implementation of the Paris Declaration on Aid effectiveness has encountered several drawbacks.

It is important to evaluate: 1. How Overseas Development Assistance

(ODA) in general has impacted on poverty reduction,

2. How the implementation of the Declaration has impacted on poverty.

Using difference-in-difference estimation technique and data from July 2003 to December 2008 obtained from 69 districts in Kenya,

4

TRENDS IN FOREIGN AID

Total Net ODA disbursements in Kenya increased from an annual average of $ 144.2 million in the 1970s to $ 544 million in the 1980s and to $688 million in the 1990s.

During the 1970’s, foreign aid averaged about 5 percent of GDP, increasing to about 8 percent in 1980’s and 1990’s.

5

TRENDS IN FOREIGN AID

6

TRENDS IN POVERTY The overall poverty has increased over the

years from a low of 40.25 percent in 1994 to 52.32 per cent in 1997 before easing to 46 per cent in 2005.

7

EMPIRICAL STRATEGY

To evaluate the impact of ODA on poverty, the following question come to mind; “What would have been the poverty levels in the districts had the ODA-funded projects not been initiated in the districts?”

Of particular interest would be answers to the following questions:

1. What is the impact of ODA on poverty in Kenya? 2. Has ODA been more effective in Kenya after the

start of the implementation of the 2005 Paris Declaration?

To answer these questions, has been used difference-in-difference methodology.

8

EMPIRICAL MODEL DIFFERENCE-IN-DIFFERENCE

35 districts with ODA funded projects as treated and 34 districts without ODA funded projects as controls.

The key assumption underlying the difference-in-difference methodology is that any selective differences between the treated and the control districts are constant over time.

Difference-in-difference methodology: The difference in average outcome in the treatment group before and after treatment minus the difference in average outcome in the control group before and after treatment.

9

THE POVERTY GAP INDEXwhere; z is the national poverty line and yi is the income per capita of district i. M is the total number of districts andN is the national population. When 1 = α, then the equation becomes;

the Poverty Gap Index (PGI) for district i.

The PGI is an indicator of the minimum cost of eliminating poverty.

The poverty line in Kenya is determined and based on the expenditure required to purchase a food basket that allows minimum nutritional requirements to be met.

10

EMPIRICAL RESULTS

The descriptive statistics: The average ODA disbursed stood at

3229698 Kenya shilling in the districts, The average income in the districts is

11584 Kenya Shillings, The average ODA expenditure in the

districts is 578211 Kenya Shillings The average district population of

484116 people.

11

TABLE 1: THE DESCRIPTIVE STATISTICS

12

EMPIRICAL RESULTS -11.When impact is felt only in the month of disbursement:

where T1 is a treatment dummy and equals one for the months when ODA funds were disbursed and zero otherwise. Xit is a vector of other factors that affect poverty.

The results show a statistically significant impact of ODA disbursements and ODA expenditures on poverty.

An increase in disbursements by 100 shillings therefore (around USD 1.42) will remove around 12 people from poverty.

Removing 12 people out of poverty would require around 1,000,000,000 Kenya shillings in ODA expenditures (around USD 14285714.29).

This implies that the amount of ODA disbursed have a higher impact on poverty than ODA expenditures.

13

EMPIRICAL RESULTS -1 The district income significantly reduces poverty. An increase income of 100,000 Kenya shillings

(around USD 1428) will remove 79 people from poverty.

ODA disbursements and expenditures both have statistically significant impacts on the intensity of poverty, implying that they both improve the welfare of the poorest of the poor.

An increase in ODA disbursements by one shilling reduces the intensity of poverty by around 75 cases.

This is a much higher impact than the reduction of 12 cases obtained as the impact of disbursements on the PGI.

ODA disbursements have stronger impact on the poorest of the poor than those who are less poor.

14

TABLE 2: IMPACT OF ODA IF IMPACT IS FELT ONLY IN THE MONTH OF DISBURSEMENT

15

EMPIRICAL RESULTS -22.When impact is felt in the month of disbursement and thereafter:

where T2 is an interaction dummy between the condition of treatment dummy treat denoted by one if the district has a project funded by ODA and zero otherwise

a time dummy denoting the time the project was initiated T2 = treat*time. In this case, the time a district gets the first funding is considered as the time of treatment.

X it is a vector of other factors that affect poverty including district income.

16

EMPIRICAL RESULTS -2

ODA disbursements significantly reduce poverty by around 12 cases for every 100 Kenya shilling.

This impact is slightly stronger than the impact obtained when it is assumed that ODA will impact on poverty only in the month of disbursement.

A one shilling increase in ODA disbursements reduces the intensity of poverty by 198.31 cases.

This is higher than the reduction of 75.77 obtained when it is assumed that the impact is only felt in the month of disbursement.

This reinforces the result that ODA disbursements have significant spillover effects beyond the month of disbursements.

17

TABLE 3: IMPACT OF ODA IF IMPACT IS FELT IN THE MONTH OF DISBURSEMENT AND

THEREAFTER

18

EMPIRICAL RESULTS -3

3.Impact of ODA after the 2005 Paris Declaration:

where Pt is a dummy variable denoting the implementation of the 2005 Paris Declaration with zero before 2005 and ones thereafter.

The implementation of the Paris Declaration has had no significant impact on poverty.

This implies that the implementation of the declaration has not improved the effectiveness of ODA.

However the implementation of the declaration has reduces the intensity of poverty by 215.61 cases

This implies that the poorest of the poor could have so far benefited from the implementation and not those who are closer to the poverty line.

19

TABLE 4: EFFECT OF PARIS DECLERATION

20

CONCLUSIONS

ODA has significantly reduced poverty in the district with the ODA-funded projects.

The impacts of ODA disbursements on poverty are found to be higher than the impact of ODA expenditures.

ODA disbursements have stronger impact on the intensity of poverty than on the Poverty Gap Index (PGI),

The ODA-funded projects have impacted the poorest of the poor more than those who are less poor.

21

CONCLUSIONS

Stronger impact on poverty if the ODA-funded projects are assumed to reduce poverty beyond the month of project implementation.

The implementation of the 2005 Paris declaration has not significantly reduced poverty as measured by PGI,

However the impact of the declaration was significant on the intensity of poverty

The implementation of the 2005 Paris declaration has several bottlenecks .

More effort therefore needs to be put in place to increase the pace of the implementation of the declaration to improve the effectiveness of ODA.

22

Thank you!